Glitch in RTC high-resolution/wake-up timer

[yo_ku_]

I am developing LPC540xxx project that uses the RTC high-resolution/wake-up timer (1kHz) as a tick.

This program often shows an abnormal value in the wake-up register. I found a description of another 1Hz RTC counter register having a glitch. There is no description about glitches in the 1kHz register, but I put the same deglitch code in RTC_GetWakeupCount() of the fsl_rtc driver as in RTC_GetSecondsTimerCount(), and it seems to work.

It seems that RTC_GetWakeupCount() needs to be modified.

[xiangjun_rong]

Hi, Yosuke,

As far as I know that the 1KHz RTC high resolution RTC counter driving clock  is an internal clock signal, there is not any external pad, how did you  deglitch the 1kHz signal?

Can you share the code of RTC_GetWakeupCount()  api function so that we can have a review?

Thanks

BR

XiangJun Rong

[yo_ku_ · Accepted solution]

Hi ,XiangJun Rong

My "glitch" is not a glitch in the clock waveform. It is that I get an abnormal value when I access an out-of-sync register on the bus.
The code is as follows, It gets the register value twice, just like RTC_GetSecondsTimerCount().

static inline uint16_t RTC_GetWakeupCount(RTC_Type *base)
{
uint32_t a, b;

/* Follow the RF document to read the RTC default seconds timer (1HZ) counter value. */
do
{
a = base->WAKE;
b = base->WAKE;
} while (a != b);

return (uint16_t)((b & RTC_WAKE_VAL_MASK) >> RTC_WAKE_VAL_SHIFT);
}

B.R. 

[xiangjun_rong]

Hi, Yosuke,

As you know that reading the base->WAKE; register is faster than counting the base->WAKE; register, so it is difficult to meet the condition while (a != b);

Hope it can help you

BR

XiangJun Rong